When patients with advanced dementia develop feeding problems their caregivers must consider treatment options of assisted oral feeding or tube feeding. Tube feeding does not result in improved survival, function or aspiration risk. Use of tube feeding continues to increase, especially for patients who are older or are African-American. Shared decision-making is the recommended model of communication for serious clinical decisions with uncertain benefit. Current clinical discussions rarely meet this standard. In a prospective study of 246 decisions to tube feed, we found surrogate decision-makers had persistently high expectations of clinical benefit, and had needs for more information and improved shared decision-making. We now propose a cluster randomized controlled trial of a previously piloted decision aid to improve decisions about feeding options for patients with advanced dementia. After randomization by site, we will enroll n=250 patients with advanced dementia and feeding problems from 12 nursing homes in North Carolina, along with their surrogate decision-makers. We project 25% African-American enrollment. Surrogate decision-makers in the intervention sites will review the decision aid and be encouraged to discuss the decision with the physician, while those in the control nursing homes receive usual care. Aim 1. To test whether a decision aid on feeding options in dementia effectively informs surrogate decision- makers in the intervention group, defined as improved knowledge of treatment options, reduced expectation of benefit from tube feeding and reduced decisional conflict. Aim 2. To test whether the decision aid improves 1 and 3 month decision-making outcomes for surrogate decision-makers in the intervention group compared to controls, defined as increased frequency of communication, increased decisional satisfaction, and decreased decisional regret. Aim 3. To test the impact of the decision aid on 1 and 3 month clinical treatment decisions. The proposed study will add to scientific understanding of shared decision-making and the effectiveness of decision aids as a method for improving the quality of decision-making and subsequent clinical treatment decisions. If successful, this intervention will demonstrate a methodology with broader application to decision-making in dementia care.